Single-hop Question Answering on Factrecall en
31.36ScoreLM-infinite
Evaluation Results
| Method | Links | |
|---|---|---|
| LM-infiniteBackbone=Mistral-7B-Instruct2025.06 | 31.36 | |
| DCSBackbone=Llama-3-8B-Instruct2025.06 | 29.89 | |
| StreamingBackbone=Mistral-7B-Instruct2025.06 | 29.64 | |
| DCSBackbone=Vicuna-7B2025.06 | 27.26 | |
| InfLLMBackbone=Mistral-7B-Instruct2025.06 | 24.64 | |
| InfLLMBackbone=Llama-3-8B-Instruct2025.06 | 19.22 | |
| InfLLMBackbone=Vicuna-7B2025.06 | 16.65 | |
| Llama-3-8B-InstructBackbone=Llama-3-8B-Instruct, Context Management=Baseline2025.06 | 15.5 | |
| StreamingBackbone=Llama-3-8B-Instruct2025.06 | 12.36 | |
| LM-infiniteBackbone=Llama-3-8B-Instruct2025.06 | 12.16 | |
| MOICEBackbone=Vicuna-7B2025.06 | 8.27 | |
| Vicuna-7BBackbone=Vicuna-7B, Context Management=Baseline2025.06 | 6.81 | |
| DCSBackbone=Mistral-7B-Instruct2025.06 | 6.64 | |
| LM-InfiniteBackbone=Vicuna-7B2025.06 | 3.3 | |
| StreamingBackbone=Vicuna-7B2025.06 | 2.74 | |
| MOICEBackbone=Mistral-7B-Instruct2025.06 | 2.64 | |
| Mistral-7B-InstructBackbone=Mistral-7B-Instruct, Context Management=Baseline2025.06 | 2.29 |